Journal of Environmental Management 113 (2012) 194e205

Contents lists available at SciVerse ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Estimating the economic value of cultural ecosystem services in an urbanizing

area using hedonic pricing
Heather A. Sander a, *, Robert G. Haight b,1
a
Conservation Biology Graduate Program, University of Minnesota, 187 McNeal Hall, 1985 Buford Avenue, St. Paul, MN 55108, USA
b
U.S. Forest Service Northern Research Station, 1992 Folwell Ave., St. Paul, MN 55108, USA

a r t i c l e i n f o a b s t r a c t

Article history: A need exists to increase both knowledge and recognition of the values associated with ecosystem
Received 20 December 2010 services and amenities. This article explores the use of hedonic pricing as a tool for eliciting these values.
Received in revised form We take a case study approach, valuing several services provided by ecosystems, namely aesthetic quality
1 August 2012
(views), access to outdoor recreation, and the benefits provided by tree cover in Dakota County, Min-
Accepted 9 August 2012
Available online 29 September 2012
nesota, USA. Our results indicate that these services are valued by local residents and that hedonic
pricing can be used to elicit at least a portion of this value. We find that many aspects of the aesthetic
environment significantly impact home sale prices. Total view area as well as the areas of some land-
Keywords:
Ecosystem services
cover types (water and lawn) in views positively influenced home sale prices while views of imper-
Economic valuation vious surfaces generally negatively influenced home sale price. Access to outdoor recreation areas
Hedonic pricing significantly and positively influenced home sale prices as did tree cover in the neighborhood
Spatial econometrics surrounding a home. These results illustrate the ability of hedonic pricing to identify partial values for
ecosystem services and amenities in a manner that is highly relevant to local and regional planning.
These values could be used to increase policy-maker and public awareness of ecosystem services and
could improve their consideration in planning and policy decisions.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction
Ecosystem services and amenities are clearly valuable, but,
because their economic values are poorly recognized, they are often
neglected in planning and policy making in the US. As a result, these
services and amenities typically decline as American communities
urbanize. Improved monetization of local and regional ecosystem
services would serve to increase their consideration in local and
regional policy making and planning, making them more difficult to
disregard.
A number of methods exist for valuing these services and
amenities. These include production function methods in which an
ecosystem service or amenity is viewed as an input into the
production of a marketed good and its value is estimated based on
that good’s price (e.g., Barbier, 2007; Barbier and Strand, 1998; Bell,
1997; Klemick, 2011; Richmond et al., 2007; Sathirathai and
Barbier, 2001; Simonit and Perrings, 2011), replacement cost
* Corresponding author. Present address: Department of Geography, University of
Iowa, Iowa City, IA 52246, USA. Tel.: þ1 319 353 2964; fax: þ1 319 353 2964.
E-mail addresses: heather-a-sander@uiowa.edu (H.A. Sander), rhaight@fs.fed.us
(R.G. Haight).
1
Tel.: þ1 651 649 5178.
analyses that use the price of the least-cost alternative means for
providing a service as a proxy for its value (e.g., Allsopp et al., 2011;
Ashendorff et al.,1997; Hougner et al., 2006; Kovacs et al., 2011; NRC,
2000), stated preference approaches that use survey results to
determine individuals’ willingness-to-pay for an increase in a service
or willingness-to-accept compensation for a decrease in a service
(e.g., Boyle et al., 1994; Campos et al., 2009; Carson et al., 1992; del
Saz-Salazar and Rausell-Köster, 2008; Earnhart, 2006; McGonagle
and Swallow, 2005; Sayadi et al., 2009), and household production
functions that value environmental goods and services based on the
sale prices of marketed goods related to them (e.g., Abdalla et al.,
1992; Bolitzer and Netusil, 2000; Brasington and Hite, 2003;
Cavailhès et al., 2009; Cho et al., 2010, 2011; Conway et al., 2010;
Geoghegan, 2002; Hardie et al., 2007; Harrington et al., 1989; Irwin,
2002; Jim and Chen, 2010; Leggett and Bockstael, 2000; Lew and
Larson, 2005; Munroe, 2007; Netusil et al., 2010; Phaneuf et al.,
1998; Poudyal et al., 2009; Siderelis et al., 1995; Snyder et al., 2008;
Tyrväinen and Miettinen, 2000; Wilson and Carpenter,1999). Each of
these methods can estimate a portion of the economic value of an
ecosystem service and is thus relevant to different applications for
which economic values for services are desirable.
We consider hedonic pricing, a household production function
approach, to provide a particularly relevant means for measuring

0301-4797/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jenvman.2012.08.031
 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205
the local and regional values of ecosystem services. Using this
method, one can elicit the economic values for different levels of
service delivery as reflected in the amount individuals pay for their
residences or related goods. As such hedonic pricing models can
help us understand an important portion of the value of ecosystem
services, the portion that contributes directly to tax bases and that
is thus particularly pertinent to community land-use policy. Using
this method, the values of ecosystem services may be estimated
straightforwardly using readily-available data. This method is also
transparent such that local and regional planners and policy
makers can readily understand how values are calculated and may
apply them to decision making.
The present study explores the use of hedonic pricing to elicit
the values of several cultural ecosystem services, ecosystem
services from which people derive nonmaterial (e.g., recreational,
educational, aesthetic, cultural) benefits (Millennium Ecosystem
Assessment, 2005) using Dakota County, a rapidly-developing
county in the Twin Cities metropolitan area (TCMA) of Minnesota,
USA, as our study area. The services examined in this study include
local aesthetic quality, access to outdoor recreation areas, and tree
cover and associated services. This research makes a number of
contributions to the existing literature:
1. Through this research, we identify the economic values resi-
dential property owners hold for the target services in this
study area thus improving our understanding of the values of
these services regionally.
2. This research provides valuable information for the evaluation
of land-use policy. Several cities in Dakota County are consid-
ering policies aimed at protecting or providing vegetation,
particularly trees, and public open space as well as preserving
the local aesthetic quality of their environment. The results of
this study will inform these policies.
3. This analysis and its results illustrate the utility and relevance
of hedonic pricing as a method for estimating the values of
ecosystem services and amenities and informing policy in
general.
4. This study provides evidence to resolve discrepancies in the
values of these ecosystem services as estimated by previous
research in this study area as well as to provide a more thor-
ough evaluation of their values through the calculation of
a single hedonic model that incorporates these multiple
services.
2. Past value estimates for ecosystem services
This study focuses on three ecosystem amenities and services of
particular concern in the study area: the provision of areas for
outdoor recreation, scenic quality, and tree cover. The first two of
these are cultural ecosystem services. The last, tree cover, is not
a service per se, but, rather provides a series of cultural, supporting,
regulating, and provisioning services, among them carbon storage,
local and regional climate regulation, enhancement of the aesthetic
environment, and air pollution mitigation (Beckett et al., 2000;
Brack, 2002; Dwyer et al., 1991, 1992; Ellis et al., 2006; Laverne and
Lewis, 1996; Laverne and Winson-Geideman, 2003; McPherson
et al., 2005; Nowak and Crane, 2002; Nowak et al., 2006a, 2006b,
2000c; Sailor, 1995; Scott et al., 1998; Simpson, 1998; Simpson
and McPherson, 1996). The present study estimates the values of
the services provided by tree cover with percent tree canopy cover
acting as a proxy for these services. All services and amenities on
which this study focuses have received some degree of attention in
the economic valuation literature and were in many cases found to
be valuable to humans, so it was expected they would be readily
195
valued in our study. However, the values estimated in these studies
vary dramatically even within the study area on which this paper
focuses. Additionally, no studies have examined all of these services
simultaneously. Thus, through this study we hoped to both resolve
issues related to service values as well as to combine them in one
hedonic model so that we might better elicit their values.
Open space, which provides many services including areas for
outdoor recreation and enhanced scenic quality, has previously
been found to contribute positively to property values. A literature
review of 30 studies on the impact of parks on residential property
values found that parks nearly always positively impacted property
values (Crompton, 2001). Although these benefits varied consid-
erably with the characteristics of parks, they were generally 10e20
percent of total property values and extended 500e2000-feet
(approximately 150e610-m) from parks. More recent studies
have supported these conclusions (Asaber and Huffman, 2009;
Bolitzer and Netusil, 2000; Cho et al., 2006, 2010; Conway et al.,
2010; Earnhart, 2006; Hobden et al., 2004; Jim and Chen, 2010;
Lutzenhiser and Netusil, 2001; MacDonald et al., 2010; Poudyal
et al., 2009; Waltert and Schläpfer, 2010; Wu et al., 2004). In
general, these studies indicate the impact of open space on prop-
erty values to be greater for natural area parks (Lutzenhiser and
Netusil, 2001; Waltert and Schläpfer, 2010), larger parks (Cho
et al., 2010; Tajima, 2003), and permanently-protected parks
(Earnhart, 2006; Geoghegan, 2002; Irwin, 2002) and that the
impact of open space on home sale price may vary with neigh-
borhood context (Cho et al., 2008, 2010; Munroe, 2007). Although
many studies have examined the value of open space, these studies
are difficult to compare due to differences in their methodologies,
study areas, and temporal coverage. However, nearly all studies
indicate a positive economic value. As such, loss or creation of open
space is likely to impact communities economically.
Scenic quality is commonly assessed by examining the charac-
teristics of views. Previous studies that assessed view quality found
that its value was often reflected in property values. A review of the
economic values associated with views found that their impact
varied, but that many view types had positive impacts on resi-
dential home values (Bourassa et al., 2004). Examination of the
studies reviewed in this article as well as subsequent studies
indicates that views of certain land-use and cover types, notably
water (Benson et al., 1998; Bishop et al., 2004; Bourassa et al., 2004;
Jim and Chen, 2006, 2009, 2010; Loomis and Feldman, 2003; Luttik,
2000; Sander and Polasky, 2009), forests (Cavailhès et al., 2009;
Tyrväinen and Miettinen, 2000), grassy areas (Des Rosiers et al.,
2002; Sander and Polasky, 2009), and urban parks (Bishop et al.,
2004; Jim and Chen, 2006), positively impact home sale prices as
do views with larger areal extents (Sander and Polasky, 2009).
Views of built and industrial land-use types may negatively impact
property values (Jim and Chen, 2009; Lake et al., 2000a, 2000b),
while views of other land-use and cover types may have little or no
impact on property values. The arrangement of features in views
has also been found to impact home sale prices (Cavailhès et al.,
2009; Cho et al., 2008). Thus, as view characteristics impact the
values of single-family homes, they are likely to impact local tax
bases. Because views are readily and irreversibly impacted by land-
use change, they should receive consideration in land-use planning
and policy making.
Tree cover in urban areas provides multiple ecosystem services
some of which, particularly the provision of local scenic quality and
climate regulation, may be capitalized in home sale prices. In
general, studies indicate that tree cover enhances home sale prices
and that impacts vary with geographic location, tree species,
landscape configuration, and tree health (Cho et al., 2009;
Dombrow et al., 2000; Holmes et al., 2006; Kovacs et al., 2011;
Mansfield et al., 2005; Morales, 1980; Morales et al., 1976; Price
196 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205

et al., 2010; Thompson et al., 1999). Studies also indicate that the
level of tree cover in a neighborhood (i.e., low versus high) impacts
the extent to which increasing tree-cover percentages impact home
sale prices and that, in areas with already high levels of tree cover,
increases in tree cover may reduce home prices (Des Rosiers et al.,
2002; Netusil et al., 2010). The impact of tree cover on home sale
prices has also been found to vary with neighborhood context. For
example, in a highly urban environment, tree cover negatively
impacted sale prices (Kestens et al., 2004). Additionally, studies
indicate that the level of tree cover within different neighborhood
areas around homes impacts sale prices, with higher levels of tree
cover in areas close to parcels (i.e., within 250-m) having relatively
high and positive impacts on home sale prices while tree cover in
more distant neighborhoods has little impact (Sander et al., 2010).
Because tree cover is readily altered by development and impacts
on home sale prices, these values could improve land-use decision-
making and policy.
Previous studies have estimated the values of some ecosystem
services in the TCMA, but their findings do not fully agree and none
have examined these services in concert. Doss and Taff (1996)
examined the impacts of wetland types and proximity on the
assessed values of residential properties in Ramsey County, MN,
a heavily-urbanized county located directly north of Dakota County
that includes the city of St. Paul. Their findings indicate that
wetland and lake proximity generally increases home values with
a move 10 m closer to a wetland or related feature increasing home
values by between $99 and $145 (1990US$), and a lake view adding
approximately $45,950 to a home’s assessed value. Anderson and
West (2006) examined the impacts of open space proximity,
type, and size on residential property values in the seven-county
TCMA that includes Dakota County. Their findings indicate that
proximity to neighborhood and specialty parks, lakes, and rivers
increases a home’s sales prices by between 0.0035% and 0.0342%
per one percent decrease in distance such that reducing the
distance from a home to a neighborhood park or specialty park
increases sales prices by about $246 and $1790 (1997US$) respec-
tively. They also found that a home’s location and neighborhood
characteristics influence the effect of open space on sales price, for
example with the impact of parks on sale prices being much higher
in more dense, wealthier neighborhoods. Because these studies
used sale and tax assessor data from a time period before the TCMA
housing boom, however, the values they estimate are difficult to
compare to values estimated based upon sales occurring during
and after this boom.
Two additional papers valued ecosystem services like those
examined in this study for a corresponding time period (2005) and
similar location (Table 1). The first of these (Sander and Polasky,
2009) focused on Ramsey County. This study valued two services
examined in this study, scenic quality and access to outdoor
recreation areas, finding that increasing access to all types of

Table 1
Marginal implicit prices calculated for ecosystem services in the metropolitan Twin Cities area of Minnesota for 2005 from three studies. Sander and Polasky (2009) was
conducted in highly urbanized Ramsey County. Sander et al. (2010) focused on both Ramsey County and an urbanizing county to the south, Dakota County. The current paper
examined Dakota County alone. Marginal implicit prices were calculated at the mean home sale price for the counties of focus: $255,955 for Ramsey County, $287,637 for
Dakota and Ramsey Counties together, and $319,073 for Dakota County alone.

Variable Sander and Polasky (2009) Sander et al. (2010) Current paper

Measurement Coeff MIP % Change Measurement Coeff MIP % Change Measurement Coeff MIP % Change
Access to outdoor recreation areas
Distance to park Road Neg $136a 0.0465 Road NS Road Neg $13a 0.0404
distance (m) distance (m) distance (m)
a
Distance to trail Euclidean Neg $119 0.0531 Euclidean NS Not used
distance (m) distance (m)
Distance to lake Euclidean Neg $216a 0.0844 Euclidean Neg $134.58a 0.0469 Euclidean Neg $129a 0.0041
distance (m) distance (m) distance (m)
Distance to stream Euclidean Neg $127a 0.0496 Not used Not used
distance (m)
Tree cover
Tree cover parcel Not used Mean % tree NS Mean % tree NS
cover cover
Tree cover in 100 m Not used Mean % tree Pos $1371c 0.4766 Mean % tree Pos $1853c 0.5807
neighborhood cover cover
Tree cover in 250 m Not used Mean % tree Pos $836c 0.2906 Mean % tree Pos $1030c 0.3228
neighborhood cover cover
c
Tree cover in 500 m Not used Mean % tree NS Mean % tree Pos $1947 0.6102
neighborhood cover cover
Tree cover in 750 m Not used Mean % tree NS Mean % tree Pos $1102c 0.3454
neighborhood cover cover
Tree cover in 1000 m not used Mean % tree NS Mean % tree NS
neighborhood cover cover
View
View area Area in m2 Pos $386b 0.1508 Area in ha Pos $213.64d 0.0743 Area in ha Pos $181d 0.0568
View richness % Neg $2834c 1.1072 Not used Not used
Grass view % of view Pos $5517c 2.1555 Not used m2 in view Pos $1741d 0.5456
Water view % of view Pos $7417c 2.8978 Not used m2 in view Pos $81d 0.0253
View of 26e50% Not used Not used m2 in view Neg $831d 0.2604
impervious surfaces
View of 51e75% Not used Not used m2 in view Neg $1,035d 0.3244
impervious surfaces
2
Forest view % of view ns Not used m in view NS
Downtown St Paul view Presence/absence Neg $11,944 4.6664 Not used
a
Per 100 m decrease.
b
Per 100 m2 (10%) increase.
c
Per 10% increase.
d
Per ha increase.
 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205
outdoor recreation areas and many view characteristics increased
home sale prices. The second study (Sander et al., 2010) focused on
both Ramsey and Dakota Counties and estimated values for tree
cover as well as recreational access. This study’s results indicated
that increased access to some types of outdoor recreation areas
increased home sale prices and that tree cover in the local neigh-
borhoods around homes was positively related to home sale price.
This study measured only view area, but found a significant and
positive relationship between this variable and home sale prices.
The values estimated by the later study (Sander et al., 2010)
generally were somewhat lower than those estimated by the earlier
one for comparable variables. The lack of agreement between these
two studies in many of their aspects as well as the lack of relevant
view and tree cover-related variables to compare between them
makes drawing general conclusions about the value of these
amenities in this area difficult. As such this study, by calculating
values for many of these amenities, will improve our understanding
of their value in this region and of how this value varies with
locational context.
3. Methods
3.1. Hedonic pricing
This study uses hedonic pricing, a household production func-
tion technique that estimates the partial economic value of changes
in an ecosystem service or amenity based on the sale prices of
similar properties (e.g., residential homes) with different levels of
that amenity, to assess the economic values of the three target
services and amenities in the year 2005. Hedonic pricing models
typically estimate the marginal implicit prices associated with
a change in the attributes of a property by estimating the rela-
tionship between these attributes and the property’s sale price or
assessed value (Freeman, 2003). Most commonly, these studies
focus on the values of single-family residential homes, but may also
utilize lease values for commercial or residential rental units. The
present study uses the sales prices of single-family residential
homes to construct a hedonic pricing model that relates sale price
to the structural, neighborhood, and environmental aspects of
homes through the use of ordinary least squares (OLS) regression
and spatially simultaneous autoregressive (SAR) error modeling.
The OLS model used may be written as:
ln Pi ¼ b0 þ b1 Si þ b2 Ni þ b3 Qi þ εi
Here, ln Pi represents the natural log of property i’s sale price, Si
represents a vector of property i’s structural characteristics (e.g., lot
size, home age), Ni is a vector of neighborhood characteristics (e.g.,
development intensity), Qi is a vector of environmental character-
istics (e.g., recreation area access, percent tree cover on parcel), and
εi is an error term. Because we expected the impact of some vari-
ables (e.g., distance variables, home square footage, lot acreage) to
decline as their values increased, we used their natural logs in our
model. We also included a squared term for home age because we
expected its impact to become insignificant or change direction at
some value, such that newer homes would decrease in value with
increasing age to a certain age and then would increase in value.
Two issues, heteroscedasticity and spatial autocorrelation, may
complicate the estimation of hedonic pricing models, and, indeed,
any model generated using OLS regression with spatial data. To
identify whether these issues complicate a particular model, one
may test for spatial autocorrelation using Moran’s I statistic. If the
result is significant, one may then use Lagrangian multiplier diag-
nostics to assess whether this autocorrelation is best explained by
assuming spatial autocorrelation in the error term (e.g., when
197
a spatially-structured predictor variable is omitted from a model);
in the lag term (e.g., when spatial autocorrelation is present only in
the dependent variable); or in both terms. Once the likely source of
spatial autocorrelation is identified, a number of modeling methods
exist to address it. One of these is SAR modeling, which augments
standard OLS models by adding a term to incorporate the spatial
structure of the autocorrelation in the dataset (Cressie, 1993;
Haining, 2003). In SAR models a user-defined spatial weights
matrix that identifies the weight of each neighbor to a given
observation is used to implement the added term. This weights
matrix may be defined in a number of ways (Anselin and Bera,
1998; Fortin and Dale, 2005), most commonly based on distance
such that closer neighbors receive higher weights in accounting for
patterns in the dependent variable not accounted for by the inde-
pendent variables.
Three types of SAR models exist: error, lag, and mixed (Anselin,
1988; Cliff and Ord, 1981; Haining, 2003). SAR error models are
used to address autocorrelation in the error term. These models add
an additional term, lWu, to the OLS expression to represent the
spatially-dependent error term’s spatial structure. The SAR error
model may be summarized as:
Y ¼ X b þ εi þ lWu
where Y is the dependent variable, X is a matrix, b is a vector that
represents the slopes associated with the explanatory variables in
the original predictor matrix, l is the spatial autoregression coef-
ficient, W represents a spatial weights matrix used in model esti-
mation, and u represents a spatially-dependent error term. The SAR
lag model, which is used to address spatial autocorrelation in the
lag term, adds a term to account for spatial autocorrelation in the
dependent variable to the standard OLS regression such that:
Yi ¼ X b þ εi þ rWY
where r is the autoregression coefficient and Y is the response
variable. When spatial autocorrelation exists in both the lag and
error terms, a SAR mixed or Durbin model may be used. This model
adds an additional term, WXg, to represent the autocorrelation
coefficient of the lagged independent variables, where g is the
autoregression coefficient. SAR mixed models may thus be repre-
sented as follows:
Yi ¼ X b þ εi þ rWY þ WX g
In implementing our hedonic pricing model, we use these
models in cases where statistical tests indicate they are
appropriate.
3.2. Study area
Dakota County, located in east-central Minnesota in the south-
eastern portion of the TCMA (Fig. 1), consists of 21 cities and 13
townships with a total population of approximately 390,000. In the
last 20 years, this county has experienced rapid population growth
and associated urbanization, changing from a largely agricultural to
a more suburban county that acts as bedroom community for the
cities of Minneapolis and Saint Paul. As calculated from a historical
land use dataset available from the Twin Cities Metropolitan
Council (TCMC), in 1984, for example, 80% of the county was
occupied by agricultural land use while 10% was occupied by built
land uses. By 2005, the year on which this study focuses, agriculture
land use had declined to occupy less than 65% of the county while
built land uses had grown to cover approximately 20%. The county
presently contains a mix of land uses ranging from urban and
suburban to agricultural and is dominated by urban and suburban
198 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205
Fig. 1. The location of the study area, Dakota County, Minnesota, USA.
land uses in the north and west and agricultural land uses in the
south and east. Rapid urbanization is expected to continue into the
foreseeable future. As such Dakota County may be considered
typical of many rapidly-urbanizing areas of the U.S.
Previous studies (Anderson and West, 2006; Sander and
Polasky, 2009; Sander et al., 2010) have considered this county as
well as other counties in the seven-county TCMA to be one housing
market that contains residential areas that support the centers of
employment found in the cities of St. Paul and Minneapolis. Home
buyers act to maximize their utility be selecting among the
neighborhood submarkets in these counties. In the present study,
we examine the Dakota County housing market in which even rural
areas over an hour’s drive from Minneapolis and St. Paul act largely
to support the Twin Cities with these cities acting as employment
centers for this county.
3.2.1. Data
Many of the data used in constructing the hedonic pricing model
originated in the Metropolitan Twin Cities Parcel Dataset, which
consists of spatially-referenced sale, tax, and structural data for all
parcels in the seven-county TCMA and is available from the TCMC.
We identified a total of 5094 single-family residential properties
that sold in Dakota County during the year 2005 for which full valid
data were available. Excluded observations included properties for
which data for one or more variables of interest were missing or
that had unlikely values for any variables, for example, indicating
that the home had been built in the year 1602 or that a 900-square-
foot home sold for over $2,000,000. These 5094 properties acted as
our sample in constructing our hedonic pricing model.
For each of our sample properties, we identified a series of
structural, neighborhood, and environmental attributes related to
property sales price (Tables 2 and 3). Most structural attributes as
well as sale prices came directly from the Twin Cities Parcel Dataset.
Structural attributes included finished square footage, lot size,
property tax rate, and home age variables as well as dummy vari-
ables to indicate sale month. The parcel dataset did not contain
some variables (e.g., numbers of bedrooms, numbers of bathrooms)
that are typically significant predictors of home sale price in
hedonic pricing studies. This creates the potential for omitted
variable bias which could impact estimates from this study if
significant variation exists within neighborhoods used as fixed
effects (see below) or the spatial weights matrix used did not
accurately identify neighbors. We addressed this using the best
available data on structural characteristics and by carefully select-
ing our spatial weights matrix and neighborhood fixed effects. In
addition to these variables, we calculated a dummy variable to
identify whether or not a home was situated in a Federal Emer-
gency Management Agency (FEMA) floodway using a GIS dataset
that delineated such floodways available from the Minnesota
Department of Natural Resources (MNDNR).
We quantified neighborhood characteristics in a GIS environ-
ment using several additional datasets. We estimated mean percent
impervious surface to quantify development intensity in a neigh-
borhood as the mean percentage of impervious surface within
 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205 199

Table 2
Definitions of variables used in the hedonic pricing model with predicted effects on home sale price.

Variable Definition Expected relationship to

home sale price
Structural variables
Price Sale price for home N/A
Acres Lot size in acres Positive
Finsqft Home finished square footage Positive
Home_age Age of home in years Positive/negativea
Tax_rate Home’s tax rate as a percentage Negative
Flood Dummy variable for location in a FEMA floodway (0 if no, 1 if yes) Negative
Neighborhood variables
Impervious Mean percent impervious surface within 500 m of home Negative
CBD Euclidean distance in meters from home to closest central business district Negativeb
(downtown Minneapolis or St. Paul)
Sale month dummy variables (ref. month is February; 1 for sale in month, 0 otherwise)
Jan Dummy variable for sale in January Positive
Mar Dummy variable for sale in February Positive
Apr Dummy variable for sale in April Positive
May Dummy variable for sale in May Positive
June Dummy variable for sale in June Positive
July Dummy variable for sale in July Positive
Aug Dummy variable for sale in August Positive
Sept Dummy variable for sale in September Positive
Oct Dummy variable for sale in October Positive
Nov Dummy variable for sale in November Positive
Dec Dummy variable for sale in December Positive
Submarket dummy variables (ref. location is Simley HSD; 1 in district, 0 otherwise)
Applevalley Dummy variable for location in Apple Valley High School district Positive
Burnsville Dummy variable for location in Burnsville High School district Positive
Eagan Dummy variable for location in Eagan High School district Positive
Eastview Dummy variable for location in Eastview High School district Positive
Farmington Dummy variable for location in Farmington High School district Positive
Hastings Dummy variable for location in Hastings High School district Positive
Lakeville Dummy variable for location in Lakeville High School district Positive
Rosemount Dummy variable for location in Rosemount High School district Positive
S_Stpaul Dummy variable for location in South St. Paul High School district Positive
W_Stpaul Dummy variable for location in West St. Paul High School district Positive
Nfld_Rndlph Dummy variable for location in Northfield or Randolph High School districts Positive
Ecosystem service and amenity variables
Access to outdoor recreation areas
Lake Euclidean distance in meters from home to closest lake Negativea
Park Road distance in meters from home to closets 1 ha or large park Negativea
Tree cover
Tree_parcel Mean percent tree cover on the home’s parcel Positive
Tree_100 Mean percent tree cover in 100 m neighborhood around parcel Positive
Tree_250 Mean percent tree cover in 100e250 m neighborhood around parcel Positive
Tree_500 Mean percent tree cover in 250e500 m neighborhood around parcel Positive
Tree_750 Mean percent tree cover in 500e750 m neighborhood around parcel Positive
Tree_1000 Mean percent tree cover in 750e1000 m neighborhood around parcel Positive
View
View_area Total areal extent of a home’s viewshed in ha Positive
IMP5_10 Area of 5e10% impervious land cover in home’s viewshed in meters Negative
IMP11_25 Area of 11e25% impervious land cover in home’s viewshed in meters Negative
IMP26_50 Area of 26e50% impervious land cover in home’s viewshed in meters Negative
IMP51_75 Area of 51e76% impervious land cover in home’s viewshed in meters Negative
IMP76_100 Area of 76e100% impervious land cover in home’s viewshed in meters Negative
Lawn Area of short grass (lawn) land cover in home’s viewshed in meters Positive
Mtd_tallgr Area of maintained tall grassland cover in home’s viewshed in meters Positive
Forest Area of forest land cover in home’s viewshed in meters Positive
Shrub Area of shrub land cover in home’s viewshed in meters Positive
Grassland Area of unmaintained grassland land cover in home’s viewshed in meters Positive
Emer_veg Area of emergent vegetation land cover in home’s viewshed in meters Positive
VW_H20 Area of open water land cover in home’s viewshed in meters Positive
Wood_wet Area of woody wetland land cover in home’s viewshed in meters Positive
AG Area of agricultural land cover in a home’s viewshed in meters Negative
a
We expected home sale price to be negatively related to home sale price to a certain age, then positive.
b
A negative relationship between distance variables and home sale price implies that home sale price decreases with increasing distance.

500-m of each property as identified by an impervious surface
percentage map for the region available from the University of
Minnesota’s Remote Sensing and Geospatial Analysis Laboratory.
Additionally, because ease of access to business centers may
influence home sale prices, we calculated distances to the central
business districts of Minneapolis and St. Paul as identified in a GIS
dataset depicting fare zones for regional transit systems available
from the TCMC. In this case, we calculated the Euclidean distance
between each property and the border of the closest central busi-
ness district. To account for the impact of the submarket in which
a home is situated on its sale price, we also generated a series of
housing submarket dummy variables based on the home’s high
200 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205

Table 3 because these areas are typically accessed using roads (Sander and
Descriptive statistics for continuous variables used in the hedonic pricing model for Polasky, 2009). As this is unlikely to be the case for properties
Dakota County in 2005.
located adjacent to or across the street from parks, we identified
Variable Mean Std. deviation Min. Max. such properties and assigned them travel distances of zero since
Structural variables their owners would likely gain access directly from their property.
Price ($) 319,073.79 141,121.28 100,000.00 2,870,250.00 Lakes also serve as significant sites for outdoor recreation in the
Acres 0.45 1.20 0.06 43.38
study region. As such, we calculated an additional metric to
Finsqft 2215.73 812.14 614.00 11,498.82
Home_age 23.50 22.54 0.00 153.00 quantify access to recreational open space, distance to lakes. In this
Tax_rate (%) 0.00 1.69 0.90 0.28 case, we identified all lakes in the study area using a dataset
Neighborhood variables available from the MNDNR. Experience and past studies conducted
Impervious (m) 28.60 18.36 0.00 100.00 in the region indicate that lakes are typically accessed at a series of
CBD (m) 22,199.30 9,267,58 1719.86 52,052.08
Ecosystem service and amenity variables
points located nearly continuously along their perimeters in an as-
Access to outdoor recreation areas the-crow-flies fashion rather than at discrete entry points (Sander
Lake (m) 898.25 824.20 0.00 9320.77 and Polasky, 2009). Thus, we calculated Euclidean distances
Park (m) 555.43 1132.93 0.00 13,095.19 between each sample residence and its closest lake to quantify lake
Tree cover
access. We use this variable with the park proximity variable to
Tree_parcel (%) 13.60 22.59 0.00 93.00
Tree_100 (%) 13.40 16.84 0.00 90.00 quantify each property’s access to outdoor recreation areas.
Tree_250 (%) 14.00 17.13 0.00 88.57 We calculated viewsheds, computational approximations of
Tree_500 (%) 14.57 17.26 0.00 90.00 views, in a GIS environment to identify the scenic quality of the
Tree_750 (%) 15.10 17.31 0.00 88.39 environment around each sample home. In so doing, we used
Tree_1000 (%) 15.80 17.73 0.00 100.00
View
several GIS datasets: a bare-earth digital elevation model (DEM)
View_area 33.26 29.83 0.66 246.58 available from the TCMC, a planimetric dataset that identified the
IMP5_10 (m2) 4487.61 13,978.91 0.00 237,225.00 footprints and locations of buildings provided by Dakota County
IMP11_25 (m2) 4422.93 18,981.18 0.00 368,825.00 GIS, and, to identify the land-cover composition of each property’s
IMP26_50 (m2) 22,979.77 32,983.27 0.00 368,625.00
viewshed, a land-cover map, Twin Cities Metro Hybrid Land Cover
IMP51_75 (m2) 6018.01 17,306.61 0.00 676,975.00
IMP76_100 (m2) 2862.58 12,844.79 0.00 454,700.00 (HYBLC) 2000, available from the MNDNR, which we updated to
Lawn (m ) 2
2583.85 11,517.40 0.00 260,225.00 2005 conditions using parcel-level land-use data. To calculate
Mtd_tallgr (m2) 2731.45 14,538.67 0.00 378,000.00 viewsheds, we first modified the bare earth DEM to include
2
Forest (m ) 6339.91 23,697.98 0.00 708,100.00 buildings, then used this DEM, along with footprints for sample
Shrub (m2) 232.17 1635.22 0.00 38,250.00
Grassland (m2) 4760.88 19,999.88 0.00 665,025.00
buildings as observer locations, to calculate each property’s
Emer_veg (m2) 2313.54 14,894.02 0.00 471,600.00 viewshed using techniques established in previous studies (Sander
VW_H20 (m2) 4904.00 33,990.69 0.00 1,162,775.00 and Manson, 2007; Sander and Polasky, 2009). The views calcu-
2
Wood_wet (m ) 2339.39 25,210.89 0.00 936,600.00 lated in this manner are based on estimated building heights and
AG (m2) 19,527.16 111,402.22 0.00 1,836,450.00
actual locations and approximate the views from top-story
windows of each building in the study. Because minimal informa-
tion regarding tree locations and heights is currently available in
school district. We selected high school districts following exami- the study area, it was not possible to include trees as view
nation of several alternative methods including elementary and obstructions in the DEM. To quantify viewshed characteristics, we
middle school districts, cities and townships, and zip codes as these next calculated the areal extent of each viewshed as well as the area
produced the lowest mean squared errors for hedonic pricing of each land-cover class in each view.
models. In this way, we identified a total of twelve housing To provide a further measure of the local aesthetic environment
submarkets. around each sample parcel as well as a measure of local climate
We estimated a series of environmental variables for each regulation, we calculated additional variables related to tree cover.
sample property, focusing on the following previously described To do so, we utilized the National Land Cover Database (NLCD) 2001
ecosystem services. To quantify each parcel’s access to outdoor Tree Canopy dataset available from the MNDNR. This dataset
recreation sites, we identified all parks of 1-ha (ha, 10,000 m2) or identifies per pixel percent tree canopy cover at a 30-m resolution
more in area using two datasets available from the TCMC, the TCMC based upon Landsat Thematic Mapper imagery (Homer et al., 2004;
Regional Recreational Open Space Features and The Lawrence Huang et al., 2003). A temporal mismatch of four years exists
Group Landmarks. Past hedonic studies have used several different between this dataset and the parcel data. However, tree cover in the
measures to quantify access to open space areas, including the size study area changes little in the short-term since most land-cover
of the closest open space area to a home (Lutzenhiser and Netusil, change is from treeless agricultural land to built land covers and
2001), dummy variables to indicate whether or not such areas tree canopy cover changes little in the short space of four years.
occur within a given distance of a home (Lutzenhiser and Netusil, Additionally, no other tree cover data were available. Thus, this
2001; Netusil, 2005), the percentage or area of land within dataset, although not ideal, was the best available dataset for our
a given buffer distance occupied by open space (Acharya and purposes. To quantify tree cover as well as to identify the sphere of
Bennett, 2001; Geoghegan, 2002; Geoghegan et al., 1997; Irwin, influence of tree cover on home sale prices, we calculated the mean
2002; Kong et al., 2007; Ready and Abdalla, 2005), and, most percent tree cover within a series of neighborhoods around each
commonly, the distance between a home and its closest open space sample parcel (parcel level, parcel to 100-m, 100- to 250-m, 250- to
area (Wu et al., 2004). Based on previous research in the study area 500-m, 500- to 750-m, and 750- to 1000-m). We expected that tree
(Anderson and West, 2006; Doss and Taff, 1996; Sander and cover in the closer neighborhoods would contribute positively to
Polasky, 2009), we chose the last measure, distance, to quantify home sale prices and that tree cover at the parcel level as well as in
each property’s open space access and calculated the road distance more distant neighborhoods would not significantly impact home
between each sample property and its closest open space area in sale prices. Although this is counter to some previous studies (e.g.,
a GIS. Past research indicates that road distance best matches Anderson and Cordell, 1988; Dombrow et al., 2000; Donovan and
residents’ perceptions of access to the large parks used in this study Butry, 2010), these studies examined only parcel-level tree cover
 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205 201

and did not include variables to control for neighborhood tree

cover. More recent studies that considered both neighborhood and
Table 4
parcel tree cover found that tree cover in local neighborhoods Results of the SAR error hedonic pricing model for Dakota County with White’s
around homes impacts home sale prices while tree cover on the standard errors.
parcel itself does not (Holmes et al., 2006; Sander et al., 2010). We
Variable Coefficient Std. error t-value
consider these studies to more accurately capture the relationship
(Intercept) 10.28700000 0.19082000 53.91 ***
between tree cover and home sale prices as they control for tree Structural variables
cover at both levels. Ln_acres 0.14003000 0.00520090 26.92 ***
The above-described variables were used to construct a hedonic Ln_finsqft 0.50105000 0.00811970 61.71 ***
pricing model as detailed in the previous section after first verifying Home_age 0.01077600 0.00029348 36.72 ***
Age_sq 0.00007356 0.00000264 27.82 ***
that variables were not correlated with one another. Based upon
Tax_rate 0.10109000 0.00895870 11.28 ***
this model, we estimated the marginal implicit prices for significant Flood 0.08327400 0.01341900 6.21 ***
environmental variables. Sale month dummy variables (ref. month is February; 1 for sale in month,
0 otherwise)
4. Results Jan 0.02610700 0.01136800 2.30 *
Mar 0.01240100 0.01167400 1.06
Apr 0.02224900 0.01144000 1.94
The results of the hedonic pricing model for Dakota County as May 0.02363600 0.01133100 2.09 *
well as the marginal implicit prices for the ecosystems services June 0.03064500 0.01088400 2.82 **
variables are presented in Tables 1 and 4. The adjusted R2 value for July 0.03810400 0.01094100 3.48 **
Aug 0.04712700 0.01105400 4.26 ***
the OLS model (0.8265) was highly significant (p < 0.001).
Sept 0.04049500 0.01137400 3.56 ***
However, because our Moran’s I estimate was significant Oct 0.03536400 0.01183800 2.99 **
(p < 0.001) and robust Lagrangian multiplier tests, used because Nov 0.03351100 0.01199200 2.79 **
simple Lagrangian lag and error tests were significant, indicated the Dec 0.02666800 0.01235300 2.16 *
presence of significant spatial autocorrelation in the error term Neighborhood variables
Impervious 0.00038039 0.00014732 2.58 *
(RLMerr ¼ 743.86, p < 0.001; RLMlag ¼ 0.35, p ¼ 0.56), we esti- LN_CBD 0.12675000 0.01894700 6.69 ***
mated a SAR error model using 2-km weights to address spatial Submarket dummy variables (ref. location is Simley HSD; 1 in district,
autocorrelation. We selected these weights carefully after exam- 0 otherwise)
ining a variety of spatial weight matrices. This model represented Applevalley 0.15034000 0.03353800 4.48 ***
Burnsville 0.09444800 0.03228000 2.93 **
an improvement over the OLS model as indicated by the significant
Eagan 0.13177000 0.03162900 4.17 ***
value for the coefficient lambda (l ¼ 0.7587, p < 0.001), the spatial Eastview 0.13064000 0.03166800 4.13 ***
autoregression coefficient, and by its reduced Akaike’s information Farmington 0.11641000 0.03608700 3.23 **
criterion value (5460) as compared to that of the OLS model Hastings 0.07867500 0.04358500 1.81
(5167.5). Because a BreuschePagan test designed for use with SAR Lakeville 0.18074000 0.03508900 5.15 ***
Rosemount 0.10153000 0.03230800 3.14 **
models indicated the presence of significant heteroscedasticity, we S_Stpaul 0.06496300 0.02492000 2.61 *
also calculated White’s standard errors (White, 1980) using W_Stpaul 0.06102800 0.02832800 2.15 *
a modified method for use with SAR models (R. Bivand personal Nfld_Rndlph 0.12062000 0.05843700 2.06 *
communication). Ecosystem service and amenity variables
Access to outdoor recreation areas
Most coefficients for structural variables were significant and of
Ln_park 0.00042724 0.00021280 2.01 *
the expected sign. The acreage of a home’s lot as well as its finished LN_lake 0.00405820 0.00062261 6.52 ***
square footage both were positively related to home sale prices, Tree cover
indicating that homes with higher acreage or finished square Mean_tree 0.00012471 0.00011822 1.05
footage sold for more than homes with less, while tax rate was Tree_100 0.00058086 0.00018516 3.14 **
Tree_250 0.00032274 0.00016329 1.98 *
negatively related, indicating that homes with higher property tax Tree_500 0.00061027 0.00016913 3.61 ***
rates sell for less than other homes. Home age was negatively Tree_750 0.00034537 0.00016138 2.14 *
related to home sale price to the age of approximately 145-years, Tree_1000 0.00007851 0.00014685 0.53
after which it was positively related. Most sale month dummy View
View_area 0.00056791 0.00009884 5.75 ***
variables, with the exception of March, had a significant or nearly
IMP5_10 0.00000010 0.00000018 0.56
significant and positive relationship to home sale prices as IMP11_25 0.00000026 0.00000014 1.81
compared to February, indicating that sale prices are significantly IMP26_50 0.00000026 0.00000008 3.42 **
higher in most other months than in February. IMP51_75 0.00000032 0.00000013 2.54 *
Location in a FEMA flood zone was positively related to home sale IMP76_100 0.00000024 0.00000017 1.44
Lawn 0.00000055 0.00000019 2.89 **
price such that homes in flood zones experienced higher sale prices
Mtd_tallgr 0.00000026 0.00000016 1.65
than other homes. This is likely a function of both the desirability of Forest 0.00000009 0.00000011 0.76
living near water and a lack of awareness of flood risk on the part of Shrub 0.00000100 0.00000126 0.79
the general public in this region. Previous studies suggest that this Grassland 0.00000007 0.00000013 0.57
Emer_veg 0.00000007 0.00000016 0.41
lack of significance may result from a lack of information related to
VW_H20 0.00000025 0.00000007 3.48 **
and a lack of understanding regarding the degree of flood risk or Wood_wet 0.00000005 0.00000010 0.52
insurance costs for properties in flood zones on the part of home AG 0.00000004 0.00000003 1.70
purchasers (Chivers and Flores, 2002; Pope, 2008). These home l ¼ 0.75867 LR test value: 294.5 p < 0.001.
buyers thus lack sufficient information to adequately consider this Significance codes: ***p ¼ 0.001, **p ¼ 0.01, *p ¼ 0.05, p ¼ 0.1.
information when negotiating purchase price. The results of an Log likelihood: 2785.984 for error model.
additional study suggest that recent experience with flooding ML residual variance (sigma squared): 0.019367 (sigma: 0.13916).
Number of observations: 5094.
increases home purchasers’ awareness of the risks and costs asso-
AIC: -5460, (AIC for OLS: -5167.5).
ciated with flooding and ensures that they are better-reflected in
home sale prices (Bin and Polasky, 2004). As the study area
202 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205
examined in the present study has little recent experience with
flooding, it is likely that homeowners were unaware of or unclear on
the risks associated with purchasing a home in a flood zone and thus
that sale prices may not accurately capture the true value associated
with this variable. Future studies might tease out a portion of this by
including variables that identify distances to rivers and streams in
addition to the distances to lakes included in this study. We
attempted to include such variables in the present study, but were
obliged to remove them because of multicollinearity.
The coefficients for nearly all neighborhood variables were
significant and of the expected sign. Distance to the central busi-
ness districts of Minneapolis and St. Paul was negatively related to
home sale price such that homes located closer to a central business
district sold for more than comparable homes located further away.
The mean area of impervious surface within 500 m of a home was
also positively related to home sale price, indicating that homes in
areas with more impervious surface and thus higher development
intensities sold for more than homes with lower levels. This result
is surprising and may indicate a preference for living in more
developed areas which might incorporate more amenities (e.g.,
restaurants, shopping and fitness centers, day care providers,
schools) or simply may be a function of the tendency of homes to be
located in more intensively-developed areas in Dakota County. This
might be perceived as indicating that increasing development is
valued, but this study did not consider very high-density forms of
residential development (e.g., townhomes, apartment buildings),
so it is difficult to comment upon this. However, as the values of
these properties tend to be lower than those of single-family
housing, one might speculate that higher intensity development
is valued only to a certain point, after which it may become a dis-
amenity. All dummy variables for submarkets with the exception of
the dummy variable for location in the Hastings High School district
were significant and positive as compared to the Simley High
School district, indicating that homes in these submarkets experi-
ence higher sale prices than those in the Simley submarket.
Both variables indicating a property’s access to outdoor recrea-
tion areas significantly impacted home sale prices. Road distance to
parks greater than 1-ha in area had a significant and negative
relationship to home sale price, such that the marginal implicit price
of a 100-m decrease in distance to such a park evaluated at the mean
home sale price of $319,073 from an initial distance of 1-km was
$13.16 (0.040%). Euclidean distance to lakes also was significantly
and negatively related to home sale price, although the impact of
lakes was greater than that of parks, with a marginal implicit price
for a 100-m decrease in distance calculated as above of $129
(0.041%). Thus, the owners of single-family properties in Dakota
County pay more to live near to these outdoor recreation areas.
The results of the hedonic pricing model also indicate that some
aspects of views significantly influence home sale prices in Dakota
County (Table 4). View area, for example, significantly and positively
impacts home sale prices such that a 1-ha (10,000-m2) increase in
view area from the mean view area (33.26-ha) calculated at the
mean home sale price corresponds to a home sale price increase of
$181 (0.057%). The areas of two built land-cover types in views, 26e
50 percent impervious surface and 51e75 percent impervious
surface, had significant and negative relationships to home sale
price, such that a 1-ha increase in each of these land-cover types
from their mean values (0.44 and 2.98-ha, respectively) resulted in a
decrease in home sale price of $831 (0.260%) and $1035 (0.324%)
respectively. The coefficients for other built land-cover types (i.e.,
5e10 percent impervious, 11e25 percent impervious, and 76e100
percent impervious) were also negative, but were generally
smaller and not significant. This indicates that the owners of
single-family homes may prefer homes with views that include
lower levels of impervious surface, below the 26 percent level.
The failure of views with very high (76e100 percent impervious)
levels of impervious surface to significantly impact home
prices may indicate that the owners of homes in highly developed
areas value something else about these areas, for instance,
their urban character, and that this offsets the negative value of
highly-developed views under other circumstances. However, the
coefficient for this variable was relatively high and negative
(0.00000024, p ¼ 0.15), suggesting a tendency on the part of home
buyers to pay less for homes with high levels of impervious surface
in their views. Additionally, in combination with the positive values
placed on increased levels of neighborhood impervious surface
described above, the negative values for many impervious land-
cover types in views may indicate a preference for living in more
intensely-developed areas, but not actually being able to see them,
for example, in situations where barriers such as slopes obstruct
views of local impervious surfaces. It may also indicate that home-
owners make a trade-off between the level of development in their
neighborhood which may provide them with access to amenities
and the level of impervious surfaces in their views.
Two other land-cover types in views, lawn and water, signifi-
cantly and positively influenced home sale prices. Evaluated at the
mean home sale price, a 1-ha increase in the area of lawn from the
mean value (2584-m2) in a home’s viewshed corresponded to a sale
price increase of $1742 (0.55%) while an equal increase in the area
of water from its mean value (4904-m2) corresponded to a sale
price increase of $81 (0.03%). This indicates a preference on the part
of single-family homeowners for views of grassy areas such as golf
courses, parks, or large-lot residential housing and a lower prefer-
ence for views of water. The areas of all other land-cover types in
views (i.e., agriculture, maintained tall grassland, forest, shrubs,
grassland, emergent vegetation, and woody wetlands) did not
significantly impact home sale prices in the study area. It should be
noted that forest land cover includes areas explicitly identified as
forest (i.e., areas of contiguous trees with no interruption by other
land cover types) and does not include, for example, urbanized land
covers with high percentages of tree cover. Thus, the lack of
a significant impact on homes sale prices for the area of forest in
viewsheds does not imply a lack of value for tree cover.
The mean percentage of tree cover in most neighborhood areas
significantly and positively influenced home prices. Notably, the
mean percentage of tree cover on the parcel itself was not signifi-
cantly related to home sale price, indicating that homeowners are
not concerned about the level of tree cover on their parcel itself.
However, the mean tree cover percentages within the 100-m, 250-
m, 500-m, and 750-m neighborhoods showed significant and
positive relationships to home sale price such that homes with
more tree cover in these areas experienced higher sale prices. The
marginal implicit prices for a 10-percent increase in tree cover
within each of these four neighborhoods from their mean values
(13.60%, 13.40%, 14.00%, and 14.57%, respectively) evaluated at the
mean home sale price were $1853 (0.581%), $1030 (0.323%), $1947
(0.610%), and $1102 (0.345%), respectively. The level of tree cover in
the 1000-m neighborhood was not significantly related to home
sale price. This indicates that, while home purchasers are not
particularly influenced by tree cover on their own parcel, they are
influenced by tree cover in its surrounding neighborhood to
a distance of approximately 750-m.
5. Discussion and conclusion
Our results clearly illustrate the utility of hedonic pricing in
eliciting the values of multiple ecosystem services. The services
examined in this study were valued by the owners of single-family
residential properties. Positive price impacts were associated with
increased view area and views of water and lawn, higher degrees of
 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205
access to outdoor recreation areas, and higher levels of neighbor-
hood tree cover. This clearly illustrates the importance of these
services to local residents who will pay more to experience greater
levels of service delivery in their neighborhoods. Land-use plans
and policies that disregard impacts on these services not only may
adversely affect them, but also are likely to negatively impact
property values and the quality-of-life of local residents.
The values estimated in this and other studies in the region for
the same time period (Sander and Polasky, 2009; Sander et al.,
2010) can help us to better understand the values of ecosystem
services and amenities in the TCMA as well as how they vary with
locational context (Table 1). All three studies indicate that
decreasing the distance between homes and lakes that suitable for
outdoor recreation increases home sale prices. The benefit is
highest ($216/100 m closer, 0.084%) in the most highly urbanized
area (Ramsey County) and is lowest ($129/100 m closer, 0.0041%) in
less heavily-developed Dakota County. This may be due to the
higher relative difficulty of accessing these features by driving in
more urban areas which makes proximity to them more valuable as
it increases walkability. The value associated with decreasing the
distance between homes and large parks differs greatly between
the two counties (by $123), indicating that urban residents value
these recreational open space features much more highly, likely
due to their scarcity and the desirability of accessing them on foot
in urban areas. Distances to trails and streams were omitted from
the current study due to multicollinearity, so we cannot draw
conclusions regarding the values of these features in the region.
The values associated with tree cover are somewhat higher
($482/10% increase in the 100-m buffer and $194/10% increase in
the 250-m buffer) in the present study than in the other study that
examined them in both Dakota and Ramsey Counties (Sander et al.,
2010) and extend to a larger neighborhood area (750-m as opposed
to 250-m). This may result from higher overall scarcity of tree cover
in more agricultural Dakota County (mean tree cover is 11.87%) as
compared to Ramsey County where mean tree cover is somewhat
higher (18.58%). This scarcity could cause tree cover to be more
highly valued in Dakota County.
Viewshed variables are difficult to compare among these studies
because of the different means used for quantifying them (i.e.,
percent versus area increase), although all studies indicate that
properties with higher overall view areas experience higher sale
prices. The two studies that examined the impacts of different land-
cover compositions in views on home sale price (Sander and
Polasky, 2009 and the present study) both found that grassy
surfaces in views were positively related to home sale prices as was
the area of water. The value of water views appears to be somewhat
lower in Dakota County than in Ramsey County, but this is difficult
to assess for the reason stated above. Both studies indicate that
forest views do not significantly influence home sale prices and
may indicate that views of more impervious surfaces (downtown
St. Paul and views of 26e75% impervious surface types) may
negatively impact home sale prices, although this clearly requires
further exploration.
Similar variation in the values of ecosystem services in urban-
ized versus suburban and rural environments may exist in other
locations. It seems likely that, as in this study, access to outdoor
recreation areas, for example, may be more highly valued in urban
areas compared to more rural areas due to the relatively higher
scarcity of open space in urban areas. Likewise, in areas like Dakota
County that are converting from predominantly agricultural to
urban land uses, tree cover is likely to be relatively scarce due to the
treeless condition of agriculture and more valuable than in more
urban locations like Ramsey County. Comparison of values among
urban, suburban, and rural environments thus warrants further
study in additional locations.
203
This research provides valuable information that may be used in
regional land-use planning. The values reported here are partial
economic values for these services (i.e., the portion accrues to
single-family homeowners). Full values are likely to be considerably
higher. Nonetheless, these partial values provide incentives to
consider these services in planning and policy making. Notably the
values calculated here, because they represent impacts on property
values and thus on community tax bases, could influence local and
regional planning and policy making and improve the consider-
ation of ecosystem services in setting such plans and policy. The
relatively low percentage impacts of these variables translate to
large values when considered for all properties within a community
and thus may dramatically impact community tax bases.
At the local and regional level, policy makers could utilize the
values calculated here and additional ecosystem service values
calculated using hedonic pricing to identify the potential economic
impacts of land-use policies. For example, they might compare the
mean distances between homes and open space areas under
present and planned land use and utilize values derived from
hedonic pricing models to identify how these changes might impact
home sale prices and tax dollars collected in their area. Additionally,
the values of these services might be used directly in policy making
aimed at maximizing property values while encouraging sustain-
able land-use practices as communities urbanize, for instance, in
identifying land-use configurations that maximize the values of
these services. Here, policies aimed at protecting services as
development occurs could be supported by these economic values.
In Dakota County, for example, the values calculated in this study
might be used to support clustered development practices which
would reduce mean lot acreage (and thus property values), but
would preserve open space which in turn would increase property
values, offsetting these negative impacts. This is supported by
similar studies that found that clustered development, while
causing somewhat lower home sale prices, would yield a higher
profit to developers via reduced construction costs (Williams and
Wise, 2008). These values might also be used to identify potential
drops in home values associated with the development of existing
open space areas. There is thus great potential to use economic
values for ecosystem services calculated using hedonic pricing
models to encourage more sustainable development.
Although hedonic pricing can elicit the values of ecosystem
services for use in land-use planning and policy, it is not a panacea
and should be used with caution. As stated above, this method elicits
the values of ecosystem services as they accrue to the owners of
single-family properties. Thus, it underestimates the values of
ecosystem services and, for services that provide value over larger
extents, these values are likely to be considerable underestimates.
As such, using estimates from hedonic pricing models may make
services appear less valuable than they are, causing these services to
be underrepresented in policies that rely upon them. As such, while
estimates of ecosystem service values generated using hedonic
pricing are clearly useful to planning and policy, they should not be
seen as comprehensive and should be augmented using additional
measures of value. These might include estimates of value from
tourism, the production of marketed goods, or estimates related to
health impacts. In incorporating such estimates, care should be
taken to avoid double-counting which would inflate the values of
double-counted services. Future studies should explore the use of
these and additional measures to generate a more comprehensive
picture of ecosystem service value and means for incorporating
value into planning and policy making at local to regional scales.
An additional word of caution related to the use of hedonic
modeling in calculating values for ecosystem services involves the
potential for omitted variable bias. In this and most hedonic pricing
studies it is impossible to account for all factors that influence
204 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205
home sale prices. The fitted model may over or underestimate the
influence of factors included in the model to compensate for the
omitted factors. As a result the values calculated for factors
included in the model may be larger or smaller than they are in
reality. For example in this study we were unable to include vari-
ables related to home condition and to numbers of bathrooms and
bedrooms. This may have influenced estimates for other variables
included in the model. The inclusion of fixed effects as well as the
use of a SAR error model may reduce this bias, but is unlikely to
eliminate it. Policy-making based upon value estimates like those
generated in this study should bear in mind that these values are
estimates based on the best available data, but may, as a result of
omitted variable bias, be somewhat too high or low.
Ecosystem services are valuable to us as humans and are integral
to maintaining our health and well-being. The lack of recognized
values for these services has facilitated their omission from plan-
ning and policy making, however, with negative consequences for
their delivery. Hedonic pricing, as illustrated in this study, provides
a means for eliciting these values in a manner that is highly relevant
to planning and policy making. As such values estimated using this
methodology have high potential to influence the sustainability of
future landscapes if used in setting land-use policy.
References
Abdalla, C.A., Roach, B.A., Epp., D.J., 1992. Valuing environmental groundwater
changes using averting expenditures: an application to groundwater contami-
nation. Land Economics 68, 163e169.
Acharya, G., Bennett, L.L., 2001. Valuing open space and land-use patterns in urban
watersheds. Journal of Real Estate Finance and Economics 22, 221e237.
Allsopp, M.H., de Lange, W.J., Veldtman, R., 2011. Valuing insect pollination services
with cost of replacement. PLoS One 3, 1e8.
Anderson, L.M., Cordell, H.K., 1988. Influence of trees on residential property-values
in Athens, Georgia (USA) e a survey based on actual sales prices. Landscape and
Urban Planning 15 (1e2), 153e164.
Anderson, S.T., West, S.E., 2006. Open space, residential property values, and spatial
context. Regional Science and Urban Economics 36, 773e789.
Asaber, P.K., Huffman, F.E., 2009. The relative impacts of trails and greenbelts on
home price. Journal of Real Estate Finance and Economics 38, 408e419.
Anselin, L., 1988. Spatial Econometrics: Methods and Models. Kluwer Academic
Publishers, Dordrecht.
Anselin, L., Bera, A.K., 1998. Spatial dependence in linear regression models with an
introduction to spatial econometrics. In: Ullah, A., Giles, D.E.A. (Eds.), Handbook
of Applied Economic Statistics. Michael Dekker, New York, pp. 237e289.
Ashendorff, C.L., Principe, M.A., Seely, A., LaDuca, J., Beckhardt, L., Faber, W.,
Mantus, J., 1997. Watershed protection for New York City’s supply. Journal of the
American Water Works Association 89, 75e88.
Barbier, E.B., Jan. 2007. Valuing ecosystem services as productive inputs. Economic
Policy, 177e229.
Barbier, E.B., Strand, I., 1998. Valuing mangrove-fishery linkages: a case study of
Campeche, Mexico. Environmental and Resource Economics 12, 151e166.
Beckett, K.P., Freer-Smith, P., Taylor, G., 2000. Effective tree species for local air-
quality management. Journal of Arboriculture 26, 12e19.
Bell, F.W.,1997. The economic value of saltwater marsh supporting marine recreational
fishing in the Southeastern United States. Ecological Economics 21, 243e254.
Benson, E.D., Hansen, J.L., Schwartz, A.L., Smersh, G.T., 1998. Pricing residential
amenities: the value of a view. Journal of Real Estate Finance and Economics 16,
55e73.
Bin, O., Polasky, S., 2004. Effects of flood hazards on property values: evidence
before and after Hurricane Floyd. Land Economics 80, 490e500.
Bishop, I.D., Lange, E., Mahbubul, A.M., 2004. Estimation of the influence of view
components on high-rise apartment pricing using a public survey and GIS
modeling. Environment and Planning B-Planning & Design 31, 439e452.
Bolitzer, B., Netusil, N.R., 2000. The impact of open spaces on property values in
Portland, Oregon. Journal of Environmental Management 59, 185e193.
Bourassa, S.C., Hoesli, M., Sun, J., 2004. What’s in a view? Environment and Planning
A 36, 1427e1450.
Boyle, K.J., Poe, G.L., Bergstrom, J.C., 1994. What do we know about groundwater
values? Preliminary implications from a meta analysis of contingent-valuation
studies. American Journal of Agricultural Economics 76, 1055e1061.
Brack, C.L., 2002. Pollution mitigation and carbon sequestration by an urban forest.
Environmental Pollution 116, S195eS200.
Brasington, D.M., Hite, D., 2003. Demand for environmental quality: a spatial
hedonic analysis. Regional Science and Urban Economics 35, 57e82.
Campos, P., Oviedo, J.L., Caparrós, A., Huntsinger, L., Coelho, I., 2009. Contingent
valuation of woodland-owner private amenities in Spain, Portugal, and Cal-
ifornia. Rangeland Ecology and Management 62, 240e252.
Carson, R.T., Mitchell, R.C., Hanemann, W.M., Kopp, R.J., Presser, S., Ruud, P.A., 1992.
A contingent valuation study of lost passive use values resulting from the Exxon
Valdez Oil Spill. Report to the Attorney General of the State of Alaska.
Cavailhès, J., Brossard, T., Foltête, J.-C., Hilal, M., Joly, D., Tourneux, F.-P., Tritz, C.,
Wavresky, P., 2009. GIS-based hedonic pricing of landscape. Environmental and
Resource Economics 44, 571e590.
Chivers, J., Flores, N.E., 2002. Market failure in information: the national flood
insurance program. Land Economics 78, 515e521.
Cho, S.-H., Bowker, J.M., Park, W.M., 2006. Measuring the contribution of water and
green space amenities to housing values: an application and comparison of
spatially weighted hedonic models. Journal of Agricultural and Resource
Economics 31, 485e507.
Cho, S.-H., Kim, S.G., Roberts, R.K., 2011. Values of environmental landscape
amenities during the 2000e2006 real estate boom and subsequent 2008
recession. Journal of Environmental Planning and Management 54, 71e79.
Cho, S.-H., Kim, S.G., Roberts, R.K., Jung, S., 2009. Amenity values of spatial
configurations of forest landscapes over space and time in the Southern
Appalachian Highlands. Ecological Economics 68, 2646e2657.
Cho, S.-H., Lambert, D.M., Roberts, R.K., Kim, S.G., 2010. Moderating urban sprawl: is
there a balance between shared open space and housing parcel size? Journal of
Economic Geography 10, 763e783.
Cho, S.-H., Poudyal, N.C., Roberts, R.K., 2008. Spatial analysis of the amenity value of
green open space. Ecological Economics 66, 403e416.
Cliff, A.D., Ord, J.K., 1981. Spatial Processes e Models and Applications. Pion Ltd,
London.
Conway, D., Li, C.Q., Wolch, J., Kahle, C., Jerrett, M., 2010. A spatial autocorrelation
approach for examining the effects of urban greenspace on residential property
values. Journal of Real Estate Finance and Economics 41, 150e169.
Cressie, N.A.C., 1993. Statistics for Spatial Data. Wiley, New York.
Crompton, J.L., 2001. The impact of parks on property values: a review of the
empirical evidence. Journal of Leisure Research 33, 1e31.
del Saz-Salazar, S., Rausel-Köster, P., 2008. The double-hurdle model of urban green
areas valuation: dealing with zero responses. Landscape and Urban Planning
84, 241e251.
Des Rosiers, F., Thériault, M., Kestens, Y., Villeneuve, P., 2002. Landscaping and house
values: an empirical investigation. Journal of Real Estate Research 23, 139e161.
Dombrow, J., Rodriguez, M., Sirmans, C.F., 2000. The market value of mature trees in
single-family housing markets. Appraisal Journal 68, 39.
Donovan, G.H., Butry, D.T., 2010. Trees in the city: valuing street trees in Portland,
Oregon. Landscape and Urban Planning 94, 77e83.
Doss, C.R., Taff, S.J., 1996. The influence of wetland type and wetland proximity on
residential property values. Journal of Agricultural and Resource Economics 21,
120e129.
Dwyer, J.E., McPherson, E.G., Schroeder, H.W., Rowntree, R.A., 1992. Assessing the
benefits and costs of the urban forest. Journal of Arboriculture 18, 227e234.
Dwyer, J.F., Schroeder, H.W., Gobster, P.H., 1991. The significance of urban trees and
forests: toward a deeper understanding of values. Journal of Arboriculture 17,
276e284.
Earnhart, D., 2006. Using contingent-pricing analysis to value open space and its
duration at residential locations. Land Economics 82, 17e35.
Ellis, C.D., Lee, S.W., Kweon, B.S., 2006. Retail land use, neighborhood satisfaction
and the urban forest: an investigation into the moderating and mediating
effects of trees and shrubs. Landscape and Urban Planning 74, 70e78.
Fortin, M.J., Dale, M.R.T., 2005. Spatial Analysis e a Guide for Ecologists. Cambridge
University Press, Cambridge.
Freeman, A.M., 2003. The Measurement of Environmental and Resource Values,
Theory and Methods, second ed. Resources for the Future, Washington, D.C.
Geoghegan, J., 2002. The value of open spaces in residential land use. Land Use
Policy 19, 91e98.
Geoghegan, J., Wainger, L.A., Bockstael, N.E., 1997. Spatial landscape indices in
a hedonic framework: an ecological economics analysis using GIS. Ecological
Economics 23, 251e264.
Haining, R., 2003. Spatial Data Analysis: Theory and Practice. Cambridge University
Press, Cambridge.
Hardie, I., Lichtenberg, E., Nickerson, C.J., 2007. Regulation, open space, and the
value of land undergoing residential subdivision. Land Economics 83, 458e474.
Harrington, W., Krupnick, A.J., Spofford Jr., W.O., 1989. The economic losses of
a waterborne disease outbreak. Journal of Urban Economics 25, 116e137.
Hobden, D.W., Laughton, G.E., Morgan, K.E., 2004. Green space borders e a tangible
benefit? Evidence from four neighbourhoods in Surrey, British Columbia, 1980e
2001. Land Use Policy 21, 129e138.
Holmes, T.P., Murphy, E.A., Bell, K.P., 2006. Exotic forest insects and residential
property values. Agricultural and Resource Economics Review 35, 155e166.
Homer, C., Huang, C., Yang, L., Wylie, B., Coan, M., 2004. Development of a 2001
national landcover database for the United States. Photogrammetric Engi-
neering and Remote Sensing 70, 829e840.
Hougner, C., Colding, J., Söderqvist, T., 2006. Economic valuation of a seed dispersal
service in the Stockholm National Urban Park, Sweden. Ecological Economics
59, 364e374.
Huang, C., Homer, C., Yang, L., 2003. Regional forest land cover characterization
using Landsat type data. In: Wulder, M., Franklin, S. (Eds.), Methods and
Applications for Remote Sensing of Forests: Concepts and Case Studies. Kluwer
Academic Publishers, Boston, pp. 398e410.
Irwin, E.G., 2002. The effects of open space on residential property values. Land
Economics 78 (4), 465e480.
 H.A. Sander, R.G. Haight / Journal of Environmental Management 113 (2012) 194e205
Jim, C.Y., Chen, W.Y., 2010. External effects of neighbourhood parks and landscape
elements on high-rise residential value. Land Use Policy 27, 662e670.
Jim, C.Y., Chen, W.Y., 2009. Value of scenic views: hedonic assessment of private
housing in Hong Kong. Landscape and Urban Planning 91, 226e234.
Jim, C.Y., Chen, W.Y., 2006. Impacts of urban environmental elements on residential
housing prices in Guangzhou (China). Landscape and Urban Planning 78, 422e434.
Kestens, Y., Thériault, M., Des Rosiers, F., 2004. The impact of surrounding land use
and vegetation on single-family house prices. Environment and Planning B 31,
539e567.
Klemick, H., 2011. Shifting cultivation, forest fallow, and externalities in ecosystem
services: evidence from the eastern Amazon. Journal of Environmental
Economics and Management 61, 95e106.
Kong, F., Yin, H., Nakagoshi, N., 2007. Using GIS and landscape metrics in the
hedonic price modeling of the amenity value of urban green space: a case study
in Jinan City, China. Landscape and Urban Planning 79, 240e252.
Kovacs, K., Václavík, T., Haight, R.G., Pang, A., Cunniffe, N.J., Gilligan, C.A.,
Meentemeyer, R.K., 2011. Predicting the economic costs and property values
losses attributed to sudden oak death damage in California (2010e2020).
Journal of Environmental Management 92, 1292e1302.
Lake, I.R., Lovett, A.A., Bateman, I.J., Day, B., 2000a. Using GIS and large-scale digital
data to implement hedonic pricing studies. International Journal of
Geographical Information Science 14, 521e541.
Lake, I.R., Lovett, A.A., Bateman, I.J., Day, B., 2000b. Improving land compensation
procedures via GIS and hedonic pricing. Environment and Planning C-
Government and Policy 18, 681e696.
Laverne, R.J., Lewis, G.M., 1996. The effect of vegetation on residential energy use in
Ann Arbor, Michigan. Journal of Arboriculture 22, 234e243.
Laverne, R.J., Winson-Geideman, K., 2003. The influence of trees and landscaping on
rental rates at office buildings. Journal of Arboriculture 29, 281e290.
Leggett, C.G., Bockstael, N.E., 2000. Evidence of the effects of water quality on
residential land prices. Journal of Environmental Economics and Management
39, 121e144.
Lew, D.K., Larson, D.M., 2005. Valuing recreation and amenities at San Diego County
beaches. Coastal Management 33, 71e86.
Loomis, J., Feldman, M., 2003. Estimating the benefits of maintaining adequate lake
levels to homeowners using the hedonic property method. Water Resources
Research 39, 1259.
Luttik, J., 2000. The value of trees, water and open space as reflected by house prices
in the Netherlands. Landscape and Urban Planning 48, 161e167.
Lutzenhiser, M., Netusil, N.R., 2001. The effect of open spaces on a home’s sale price.
Contemporary Economic Policy 19, 291e298.
MacDonald, D.H., Crossman, N.D., Mahmoudi, P., Taylor, L.O., Summers, D.M.,
Boxall, P.C., 2010. The value of public and private green spaces under water
restrictions. Landscape and Urban Planning 95, 192e200.
Mansfield, C., Pattanayak, S.K., McDow, W., McDonald, R., Halpin, P., 2005. Shades of
green: measuring the value of urban forests in the housing market. Journal of
Forest Economics 11, 177e199.
McGonagle, M.P., Swallow, S.K., 2005. Open space and public access: a contingent
choice application to coastal preservation. Land Economics 81, 477e495.
McPherson, G., Simpson, J.R., Peper, P.J., Maco, S.E., Xiao, Q.F., 2005. Municipal forest
benefits and costs in five US cities. Journal of Forestry 103, 411e416.
Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being:
Current State and Trends Findings of the Condition and Trends Working
Group. Island Press, Washington, D.C.
Morales, D.J., 1980. The contribution of trees to residential property value. Journal of
Arboriculture 6, 305e308.
Morales, D.J., Micha, F.R., Weber, R.L., 1976. The contribution of trees to residential
property value: Manchester, Connecticut. Valuation 23, 26e43.
Munroe, D.K., 2007. Exploring the determinants of spatial pattern in residential land
markets: amenities and disamenities in Charlotte, NC, USA. Environment and
Planning B: Planning and Design 34, 336e354.
NRC, 2000. Watershed Management for a Potable Water Supply: Assessing the New
York City Strategy. National Academy Press, Washington, D.C.
Netusil, N.R., 2005. The effect of environmental zoning and amenities on property
values: Portland, Oregon. Land Economics 81, 227e246.
Netusil, N.R., Chattopadhyay, Kovacs, K.F., 2010. Estimating the demand for tree
canopy: a second-stage hedonic price analysis in Portland, Oregon. Land
Economics 86, 281e293.
Nowak, D.J., Crane, D.E., 2002. Carbon storage and sequestration by urban trees in
the USA. Environmental Pollution 116, 381e389.
Nowak, D.J., Crane, D.E., Stevens, J.C., 2006a. Air pollution removal by urban trees
and shrubs in the United States. Urban Forestry and Urban Greening 4, 115e123.
Nowak, D.J., Hoehn III, R.E., Crane, D.E., Stevens, J.C., Walton, J.T., 2006b. Assessing
Urban Forest Effects and Values, Washington, D.C.’s Urban Forest. Northern
205
Research Station, Resource Bulletin NRS-1. U.S. Department of Agriculture,
Forest Service, Newtown Square, PA.
Nowak, D.J., Hoehn III, R.E., Crane, D.E., Stevens, J.C., Walton, J.T., 2006c. Assessing
Urban Forest Effects and Values, Minneapolis’ Urban Forest. Northeastern
Research Station, Resource Bulletin NE-166. U.S. Department of Agriculture,
Forest Service, Newtown Square, PA.
Phaneuf, G.J., Kling, C.L., Herriges, J.A., 1998. Valuing water quality improvements
using revealed preference methods when corner solutions are present. Amer-
ican Journal of Agricultural Economics 80, 1025e1031.
Pope, J., 2008. Do seller disclosures affect property values? Buyer information and
the hedonic model. Land Economics 84, 551e572.
Poudyal, N.C., Hodges, D.G., Ronn, B., Cho, S.-H., 2009. Valuing diversity and spatial
pattern of open space plots in urban neighborhoods. Forest Policy and
Economics 11, 194e201.
Price, J.I., McCollum, D.W., Berrens, R.P., 2010. Insect infestation and residential
property values a hedonic analysis of the mountain pine beetle epidemic. Forest
Policy and Economics 12, 415e422.
Ready, R.C., Abdalla, C.W., 2005. The amenity and disamenity impacts of agriculture:
estimated from a hedonic pricing model. American Journal of Agricultural
Economics 87, 314e326.
Richmond, A., Kaufmann, R.K., Myneni, R.B., 2007. Valuing ecosystem services:
a shadow price for net primary production. Ecological Economics 64,
454e462.
Sailor, D.J., 1995. Simulated urban climate response to modifications in surface
albedo and vegetative cover. Journal of Applied Meteorology 34, 1694e1704.
Sander, H.A., Manson, S.M., 2007. Heights and location of artificial structures in
viewshed calculation: how close is close enough? Landscape and Urban Plan-
ning 82, 257e270.
Sander, H.A., Polasky, S., 2009. The value of views and open space: estimates from
a hedonic pricing model for Ramsey County, Minnesota, USA. Land Use Policy
26, 837e845.
Sander, H.A., Polasky, S., Haight, R.G., 2010. The value of urban tree cover: a hedonic
property price model in Ramsey and Dakota County, Minnesota, USA. Ecological
Economics 69, 1646e1656.
Sathirathai, S., Barbier, E.B., 2001. Valuing mangrove conservation, southern
Thailand. Contemporary Economic Policy 19, 109e122.
Sayadi, S., González-Roa, M.C., Calatrava-Requena, J., 2009. Public preferences for
landscape features: the case of agricultural landscapes in mountainous Medi-
terranean areas. Land Use Policy 26, 334e344.
Scott, K.I., McPherson, E.G., Simpson, J.R., 1998. Air pollution uptake by Sacra-
mento’s urban forest. Journal of Arboriculture 24, 224e234.
Siderelis, C., Brothers, G., Rea, P., 1995. A boating choice model for the valuation of
lake access. Journal of Leisure Research 27, 264e282.
Simonit, S., Perrings, C., 2011. Sustainability and the value of the ‘regulating’
services: wetland and water quality in Lake Victoria. Ecological Economics 70,
1189e1199.
Simpson, J.R., 1998. Urban forest impacts on regional heating and cooling
energy use: Sacramento County case study. Journal of Arboriculture 24,
201e214.
Simpson, J.R., McPherson, E.G., 1996. Potential of tree shade for reducing residential
energy use in California. Journal of Arboriculture 22, 10e18.
Snyder, S.A., Kilgore, M.A., Hudson, R., Donnay, J., 2008. Influence of purchaser
perceptions and intentions on price for forest land parcels: a hedonic pricing
approach. Journal of Forest Economics 14, 46e72.
Tajima, K., 2003. New estimates of the demand for urban green space: implications
for valuing the environmental benefits of Boston’s big dig project. Journal of
Urban Affairs 25, 641e655.
Thompson, R., Hanna, R., Noel, J., Piirto, D., 1999. Valuation of tree aesthetics on
small urban-interface properties. Journal of Arboriculture 25, 225e234.
Tyrväinen, L., Miettinen, A., 2000. Property prices and urban forest amenities.
Journal of Environmental Economics and Management 39, 205e223.
Waltert, F., Schläpfer, F., 2010. Landscape amenities and local development:
a review of migration, regional economic and hedonic pricing studies. Ecolog-
ical Economics 70, 141e152.
White, H., 1980. A heteroskedasticity-consistent covariance matrix and a direct test
for heteroskedasticity. Econometrica 48, 817e838.
Williams, E.S., Wise, W.R., 2008. Economic impacts of alternative approaches to
storm-water management and land development. Journal of Water Resources
Planning and Management 135, 527e546.
Wilson, M.A., Carpenter, S.R., 1999. Economic valuation of freshwater ecosys-
tems services in the United States: 1971e1997. Ecological Applications 9,
772e783.
Wu, J.J., Adams, R.M., Plantinga, A.J., 2004. Amenities in an urban equilibrium model:
residential development in Portland, Oregon. Land Economics 80, 19e32.